Application and Research of Multi-Pose Face Recognition Method Based on Intelligent Classroom

This study aims to improve the accuracy and reliability of multi-pose face recognition in intelligent classroom. In educational scenarios, facial recognition faces challenges such as pose change, illumination difference and partial occlusion. Therefore, this paper proposes a method that combines the improved RetinaFace model with the generation of adversarial network (GAN). Firstly, MobileNet v3 and attention mechanism are used to optimize the RetinaFace model to improve the face detection accuracy. Secondly, GAN is used for positive processing of face images with different poses to ensure that the generated images are realistic and retain identity features. The experimental results show that the proposed method performs well under a variety of complex conditions, especially under extreme attitude changes, and maintains a high recognition rate. The research provides an efficient and robust face recognition solution for smart classroom and promotes the development of intelligent education.